Table of Contents
Advertisement
LBXXXA Series USB PowerSensor+™
Product Manual PMA (Power Meter Application)
Measurement Example: OFDM Pilot Tone (LB4XXA/LB6XXA models only)
Pilot tones for some wireless devices can operate at very low PRF (well below 100Hz); very low duty cycles (<0.05%);
and at a variety of carrier frequencies. Making peak power or pulse power measurements under these circumstances
can be difficult without expensive equipment. However, the LB4XXA series of sensors are quite capable of making
these measurements. In addition, the LB480A can present a time domain representation of the pulses. For this
example (a wireless LAN) we will assume a frequency of 2.4 GHz; a PRF of about 10 Hz; and a signal level of 0dBm
to 20dBm.
Assuming the power level is less than +20 dBm we can connect the sensor directly to the output of our device without
fear of damage. If your device produces power levels greater than +20 dBm you will need to add a pad between the
sensor and the output of your device. You may find it useful to enter offset or response correction if you add an
attenuator.
Press the Preset button on the power panel and configure the sensor as follows:
Mode = Pulse
Frequency = 2.4 GHz
Averages = 75 (default)
Anti-aliasing = OFF (default)
Extended Averaging = OFF (default)
Max Hold = Off (default)
Triggering = Internal, Continuous (default)
We happen to know in our case that this pilot tone has a PRF of about 9.7 Hz and a duty cycle of less than 1%. With
these settings the measurement is likely to be very unstable. It may bounce wildly between very low levels and very
high levels. The reason for this is the following:
With a PRF of about 10 Hz we get a pulse about every 100 msec. However, we have the averaging set to 75. Since
each average takes about 250 usec, a total of 75 averages results in a measurement time of about 20 msec. The
probability of intercept (between measurement time and pulse) is about 20%.
This low probability of intercept causes the bouncing. This bouncing occurs because the number of samples (75) does
not measure the signal over a long enough period of time. We could extend the averages to just enough to cover the
period or about 375 averages. This will give us much more stable readings. However, they are just on the verge of
being unstable. To resolve this problem we need to extend the averaging so that we get 10 entire cycles to measure
for 1 second. A quick calculation yields:
Averages = (1 second /250usec) = 4000
You should see a dramatic stabilization of the readings. In our example we get the following readings:
Pulse (main reading): 16.9 dBm (+/-0.2 dB)
DC = 0.05% (+/-0.05%)
Pk = 18.4 dBm (+/0.2 dB)
Avg = -14.4 dBm (+/-0.4 dBm)
While this is a dramatic improvement it is still not as stable as it could be. We can take two approaches to further
stabilize the measurements. The first approach is to simply increase the averaging to 10000.
Averages = 10000
Revision: 08/02/11
- 22 -
Advertisement
Table of Contents
